From Early Startups to Product Leadership with Mike Agnich | Ep. 12

Tim Berglund talks to Mike Agnich (Confluent) about his career in product leadership and startups. Mike’s first job: refereeing youth basketball. His challenge: leading product across connectors, governance, stream processing, and partnerships at Confluent.

SEASON 2
Hosted by Tim Berglund, Adi Polak and Viktor Gamov
Produced and Edited by Noelle Gallagher, Peter Furia and Nurie Mohamed
Music by Coastal Kites 
Artwork by Phil Vo 

•  🎧 Subscribe to Confluent Developer wherever you listen to podcasts. 
• ▶️ Subscribe on YouTube, and hit the 🔔 to catch new episodes.
• 👍 If you enjoyed this, please leave us a rating. 
• 🎧 Confluent also has a podcast for tech leaders: "Life Is But A Stream" hosted by our friend, Joseph Morais.

Transcript

SPEAKER_01 0:00

Today, from refereeing to product leadership, maybe they're the same thing. This is Confluent Developer.

SPEAKER_00 0:06

Yeah, as a referee, at least you can eject the coach if you need to.

SPEAKER_01 0:09

Yeah. You can't do that as a No, you can't. You just have to deal with it.

SPEAKER_00 0:12

So that's so the rest of my career is really starting companies. That's kind of what I did. I think the lesson I took from that is like the starting a business and growing it is a lot of doing whatever has to get done. And you can have a big impact if you're willing to kind of step out of your comfort zone and do that stuff.

SPEAKER_01 0:27

Hello there, everyone. I'm Tim Berglund and welcome to Confluent Developer, where we explore the fascinating journeys of software developers tackling hard problems. In this episode, I'm interviewing my colleague Mike Agnich. From his humble beginnings as a referee of middle school boys' sports as a high school boy to becoming a senior product leader in my favorite data streaming platform vendor, Confluent, Mike shares the story of his job as a college intern doing real-time video processing in the days before it was a little easier, and doing it in a way to make American football a little easier to visualize on TV. Let's get to it. Thanks, Tim. It's great to be here. Good to have you on the show. Um hey, tell uh in case anybody doesn't know, who are you? What do you do?

SPEAKER_00 1:23

Uh I've been a product leader here at Confluent for just over seven years. You just had an anniversary, I think. I just did. I just did. I got you know, you get all these you have these days and your LinkedIn blows up and everybody's like, hey, congratulations. Like, oh, is that now? Yeah. Just pass it.

SPEAKER_01 1:36

The people that message you then and on your birthday, but you don't really get it. That's it. It's just super awkward. Birthday and work anniversaries. That's exactly right. It's it's kindness.

SPEAKER_00 1:45

No, it's nice. So that it's uh yeah, I've been here for uh just over seven years now and been in a number of different product roles. Uh these days I lead kind of the portfolio of products that doesn't include Kafka, is the simplest way to put it. So the the connectors and what we're doing with governance and um spend a lot of time on in in stream processing and table flow with unified storage. There's a developer experience part of what my team does and a partnerships angle. So I've got my my hands in a lot of cookie jars.

SPEAKER_01 2:17

Yeah, you do. Good cookie jars too. It's fun. And you better believe there will be links in the show notes. So um how did you start? What was your very first job? Like not technical job, I mean like paperboy.

SPEAKER_00 2:29

Well, non non-technical job was I was uh I played a lot of sports. And so um outside of camp counselor, which was probably the first thing I got paid to do. Oh, nice. Uh, you know, those summer things, yeah, I did I refereed and coached basketball. Okay. So I got paid to do that, which was a lot of fun. And I learned like refereeing is a lot harder than it looks. Like if you are that job is a very thankless job. I'm sure it is. You're you're obviously if you're uh if you're a ninth grader refereeing a middle school basketball game, I think parents have strong opinions. Yes. And you're really not very good at your job. And uh uh inadvertent whistles happen all the time. It's a it's a giant mess, but that would be the first consistent thing I had was doing some like refereeing and tournament stuff.

SPEAKER_01 3:16

Yeah, which is a job where you have to make difficult calls and a substantial fraction of your constituency thinks you're an idiot.

SPEAKER_00 3:25

Yes, they all do all the time.

SPEAKER_01 3:27

Yes. Um I have always said uh that there are maybe many reasons why I'm not in product. Um but I think something about product management is really that that you have to make these difficult really prioritization calls. You're trying to sense, okay, what do people want? Here's what people are asking for, how what how do we prioritize engineering resources? No matter what, I don't know if it's a majority, but a significant plurality of your constituency thinks you're an idiot all the time. I mean, I don't think that about you, but you as a PM, a bunch of people just think you're making the wrong calls. Yeah.

SPEAKER_00 4:03

I I think that's true. The difference is that those people are usually slightly more mature than six to eighth grade boys. Which I think, like as a class of people, is like the worst class of people. Honestly. Six to eighth grade boys there are um annoying. Uh-huh. As a person, as somebody who was one at one point, that's a tough group.

SPEAKER_01 4:22

That's a tough group. Okay. So hopefully, like, you know, folks at the field are not that. No, no. Okay. I that anyway, I'd never thought that before. It just occurred to me that, okay, referee, PM, similar kinds of yeah, you had to develop some thickness to your skin. And that's really I, you know, for myself, like why am I not in product? Like maybe I'm not qualified. I don't know, whatever. But I just wouldn't be able to stand everybody thinking I'm stupid all the time. Like I have to do this thing where people like what I do.

SPEAKER_00 4:47

Yeah, as a referee, at least you can eject the coach if you need to. Yeah.

SPEAKER_01 4:50

You can't do that as a PM. No, you can't. You should have to deal with it. Um okay, so from ref, that led to some other series of steps and you need to what you're doing now. Um, whether it's here at Confluent or just whatever in your career, what's what's maybe the hardest or just most interesting problem you've solved? Now a quick word from our sponsor. Confluent Developer the Podcast is brought to you by Confluent Developer the website, which has everything you need as a developer of data streaming systems. And it's completely free. We've got curriculum, hands-on exercises, executable tutorials, the online data streaming engineer certification, also free, a way to find a meetup near you, those are free. Everything is there. I really want you to be successful in your journey as a data streaming engineer, and this is the site that has what you need. Check it out at developer.confluent.io. That's developer.confluent.io. Now back to the show.

SPEAKER_00 5:47

So I had a very cool internship when I was undergrad. And this was my first, I would say, full-time paid to write code. And I'm really not sure they got their money's worth, but I was probably after my sophomore year of undergrad.

SPEAKER_01 6:03

I worked for exactly the same soft sophomore year, January, got a job writing software. So good.

SPEAKER_00 6:08

So my job was at a startup, and the startup was called Sport Vision. Okay. And what they did, they were one of the companies, there were two at the time, that did like the yellow line for NFL games. And uh, so that first down line. Right. They also did like the hockey puck thing, superimposed in real time on a video stream.

SPEAKER_01 6:28

Yes. Okay.

SPEAKER_00 6:28

And and I worked, uh, and I, you know, I wrote some code like I said. If they used any of it, I'd be surprised. I really am not sure they early in your career, you got to be. I'm really not sure they were happy with my employment there. I can't, I can't say that they were for sure. But I got to work on a NASCAR thing they did. And um, but what was interesting, I mean, this was this was late 90s, so far before the idea of cloud infrastructure. So the way this product was built was servers that saw that sat in the back of an 18-wheeler and they had to drive these things around from stadium to stadium and run them locally. And so it was like a bunch of color keying technology that could tell the difference between a Green Bay Packers jersey and the green field uh and and the color of the field below it. Okay. And and actually in real time kind of superimpose the line with the curvature of the field. Yeah. Very cool technology. And uh, you know, I think that what I loved about that was uh while I may not have been very good at writing production code at the time, as a uh, you know, whatever I was, a 20-year-old kid, I was pretty excited about working the games. Whereas I think most of the people would rather stay home on the weekends. Uh I accepted all the invitations to go and work. The college games work, the NFL games.

SPEAKER_01 7:42

So you had an athletic background with sports replies.

SPEAKER_00 7:47

It's not like you don't picture me as the person in the truck actually doing important things. They would usually send me with binoculars up into the catwalks of the stadium, and I would up there, be up there like spotting the ball. I would say the ball's on the 27.4 yard line, that kind of stuff.

SPEAKER_01 8:03

Okay.

SPEAKER_00 8:04

And but it was a very interesting technical challenge, right? Because you had these cameras that had to actually triangulate. You had a you had a 3D map of the field, and you had these, your points of reference were the cameras.

SPEAKER_01 8:14

Okay.

SPEAKER_00 8:15

And so you had to be able to convert a camera angle to a line, a not like a non-linear line. It was basically a curved line that would move up and down the field. And it was doing real-time color keying. And as the game went on, sometimes the fields would get muddier, the shadows would change. So you had to go a day early and do all this mapping and like track the sun throughout the day so that you could be sure that if the sun went down, it would continue to work. What's the curved line?

SPEAKER_01 8:41

Help me out.

SPEAKER_00 8:42

So fields, football fields, especially, they're not flat. They have a crown because they want the water to run off to the sides.

SPEAKER_01 8:49

Certainly.

SPEAKER_00 8:50

And so they are not flat and not all fields are the same. And uh uh so you kind of have to go create those maps and correct it and then get ready to do it. So I would say being able to work around that problem, I loved that because it was such like a great real world versus theoretical world thing, and you actually got to go and see and run it and learned a lot. I got way more out of that than than they got, I think. But yeah, I think uh loved loved the startup thing. Like I kind of got hooked on the startup thing there uh and you know, saw what it meant. They that was probably uh in terms of the people building the technology, it was probably about 30 people. Uh, and it was it was an amazing, an amazing summer. And I uh I wish I could have been more helpful to them, I think. But I did, I did, I did travel to a lot of games and and and do some work.

SPEAKER_01 9:39

I want to go track down some of the people you worked with and see if number one, if they remember you, yeah, and if their assessment of you is as negative as your own. I I I think you should be a little kinder to 20-year-old Mike. Yeah. Um you know, he's out there doing his best. I I was doing my I was doing my best. Proud of that kid. Um and it's interesting. I mean, I'm just I'm I'm I'm I'm listening to you. Their cameras, were they at a fixed angle?

SPEAKER_00 10:06

Obviously, every stadium is different, but there was like one center and two on the sides, right? So you can go, if anybody goes to a game, you can kind of spot the three cameras. I think it roughly works the same way.

SPEAKER_01 10:15

Okay.

SPEAKER_00 10:16

Uh and so you had to, each of those cameras had to be able to map. And so you figure what its two-dimensional view is, you've got to map that to a 3D reality and then draw the curve over.

SPEAKER_01 10:25

And you were at some point in your undergraduate program, you know, you're gonna learn the linear algebra involved in you know, the matrix transforms that that are are there, okay, fine. Um but like you're saying, things get muddy, the shadows change, and you actually have to engineer a system that people trip on cables. Yeah, people tricked. So is it still gonna work? And you have to what um from a technology standpoint, just out of my curiosity, what is this what language, what how is this working?

SPEAKER_00 10:51

Well, um that's a good question. I remember that we were on we did one of the cool things is we had those um those SGI workstations. Okay. So that was a big deal back then that you got to work on one of those things. And actually, their office was next door to SGI at the time, which is now this like computer science history museum. Computer history museum.

SPEAKER_01 11:11

It's it's that's 10 minutes away from where we are right now. That's right. Yeah. And uh shameless plug. If you're ever in Mountain View in Bay Area, I love that place. I love it. I make I weep every time I'm there. So anyway, but back to SGI, uh SGI workstations. So I'm gonna give it a nice.

SPEAKER_00 11:26

Yeah, that was that was C. Yeah. That was C. And and because I remember though there was a lot of uh a lot of memory allocation that had to be done and deallocation like that.

SPEAKER_01 11:35

They're for memory freeing. That's hopefully you did right and most of the time. Yes.

SPEAKER_00 11:39

Yes. So that was that was C. Uh and uh uh but I said I think the first job they gave me was to get the printer working in the lobby. So so it was, I don't want to like over. I I did I did write code that summer, but uh but I also was was the person who could just make the office run better if that's what they needed.

SPEAKER_01 12:01

Yeah, yeah, of course. You're a kid, you're you're uh you're gonna do whatever you have to. So what did that how did it change you? I mean, it you remember it still some number of years later. Um you were in a computer science program?

SPEAKER_00 12:14

Computer science, I was at Stanford.

SPEAKER_01 12:15

You're Stanford, okay. Also nearby. Yep. Um and like what did that do to you in your academic program, you're thinking about your career. How did it how did it shape you?

SPEAKER_00 12:26

Well, I mean, it was very eye-opening uh like getting out of the computer lab and getting into an actual you know working code base that had to be deployed on these things that had to drive around the country. That was a whole then they didn't teach us really any of that stuff in school, which is why you feel so useless. You get in there and you're like, okay, I know how to write code, but that seems like it's only about five percent of actually what is happening here. And it's uh so that was very eye-opening and why again, why I felt like I I wasn't contributing as much as I should, maybe at times.

SPEAKER_01 12:57

Yeah, plus your sophomore year, you're like, I just I yeah, I here's a linked list. Right.

SPEAKER_00 13:01

You know, that's right. And and it's uh and it's a startup too, so it's not like you've got it's not you were showing up from some giant intern program. I was the only intern. Right. And there wasn't, you know, uh infrastructure set up to like make me successful. I kind of had to figure it out. Right.

SPEAKER_01 13:17

Uh what a great test that is, though, because that then you know you have that experience, and you know, then you got into startups and other things in your life where again, sometimes there's not infrastructure and you just need to figure it out. That's right.

SPEAKER_00 13:30

Yeah. And so I think that's the so that's so the you know the rest of my career was really starting companies. That's kind of what I did.

SPEAKER_01 13:35

Okay.

SPEAKER_00 13:36

Bonus question.

SPEAKER_01 13:37

Uh favorite company starting experience.

SPEAKER_00 13:42

Uh what the Well I started I started three, really. I loved all of them. So the first one, actually, a couple of folks that I went to school with, they had started working on this company, and and this company is Azzle, which is like still around, awesome company, awesome technology. They're all many of them are still there. But that was kind of the first experience coming out of undergrad and joining the thing that they had started to work on and was spent most of a decade there building that. And so that was awesome because it was friends and coworkers, which is both like very awesome and very unhealthy at the same time. I would say there were a lot of very long hours in that time. But uh, you know, that was a very interesting time in history. That was basically 2000, 2001. And the internet was still brand new. And I think the secret to success then was just to run as fast as you could into open space. And land grab. It was a it was a land grab. And so there was an advantage having a bunch of people who knew each other really well and effectively had absolutely no boundaries on work-life balance, was actually a pretty big competitive. Young people, maybe without a lot of other responsibilities in life. And that's and so that's that's kind of what we did. And so it was, you know, there absolutely loved that time, learned a lot. You know, it was one of these things where the code that I wrote there at first was actually they've got kind of a manufacturing and a distribution part of their business. And I wrote code to like automate the printing of things and the packaging of things and the shipping of things. And that, I guess, qualified me to run actually manufacturing and distribution, you know, as a, I don't know, 22-year-old. Sure. And so, you know, ended up hiring dozens and then probably at one point, uh, probably over a hundred people as as maybe somebody in my mid-20s, I guess. Wow. Which um you learn a lot very quickly.

SPEAKER_01 15:31

You learn a lot. It was a lot of responsibility. You're not gonna get everything right, but you you have a lot of chances to grow.

SPEAKER_00 15:36

And especially there, you know, the the nature of that, like it was just um you're not hiring all those people I was hiring, these were not like computer science grads. It was a bunch of people that had to operate stuff at scale. So, so multiple shifts a day and um a very different something nobody trained me to do. So learned a lot there. Again, kind of um just had to, I think the lesson I took from that is like the starting a business and growing it is a lot of doing whatever has to get done.

SPEAKER_01 16:08

Yeah.

SPEAKER_00 16:09

And so if that doesn't match the job description you were hired to do, then that's totally fine. And you can have a big impact if you're willing to kind of step out of your comfort zone and do that stuff. So I love that. Uh second company I started was uh actually with another classmate of mine when I was at grad school, a company called Predictify, which was like um uh in a sense, like a it was a it was we were really building an information market. So you could come and really ask a deterministic question, you could get answers, and and it would give you a ton of interesting data and then track over time if you were right or wrong. Like some of these more modern prediction markets that have ex that have come become popular in the recent days, I would say we were a very early version of that. Okay.

SPEAKER_01 16:51

Okay, prediction market if for the listeners like uh almost like you can place a bet on what you think is gonna happen. Yeah, that's right and the average price uh that's right. It uses basically a pricing mechanism to aggregate the opinion of a bunch of people. That's right. They tend to be super accurate in producing or predicting like elections and things like that. Fascinating thing. Yeah.

SPEAKER_00 17:09

And uh and so I did that for for a few years, but I would say we were a um a casualty of the 2009 downturn. Sure. And uh so that one, uh built some amazing stuff, the team was fantastic, didn't work out to where we we hoped it would go. Sure. And then uh the last company I started was a company called Terrain, which ended up being acquired here in Deconfluent, which is how I came here. And that's where I kind of made the jump over to Enterprise. And we had built a product that you know leveraged leveraged some Kafka and and Elastic and some other stuff on the back end to do what really we were aiming uh and working a lot of like search and recommendation systems, but it ended up being there was a lot of stuff there about real-time ETL that had to get done, getting all the data into one place and uh being able to like visually organize it and work with it. Uh and uh and that company ended up getting acquired into Conflo, which is how I came in here.

SPEAKER_01 18:04

And the rest is history. A lot of a lot more to talk about on future episodes. I'd like to dig into some of those things uh some more in the future. I would love to do that. All right. My guest today has been Mike Agnich. Mike, thanks for being a part of Conflo Developer. Thanks, Ted.